Trial Design & Biostatistics
This section covers the methodological and statistical decisions that shape a clinical trial's structure before a single subject is enrolled: how the trial is designed, what population it targets, what endpoints it measures, and how its results will be analyzed. These decisions are foundational in the sense that most other operational and regulatory activity in a trial follows from them, which is why this cluster is organized separately from conduct-focused and compliance-focused topics elsewhere in the site. ICH E8(R1), General Considerations for Clinical Trials, provides the overarching framework for what constitutes a well-designed trial, emphasizing quality by design and the identification of factors critical to a trial's ability to answer its research question reliably. It addresses matters such as trial objectives, population selection, and the general categories of trial design (for example, exploratory versus confirmatory) without prescribing specific statistical methods. ICH E9, Statistical Principles for Clinical Trials, is the companion guideline that addresses the statistical substance: hypothesis formulation, randomization and blinding, sample size determination, control of bias, handling of missing data, and the pre-specification of analysis plans needed to ensure a trial's conclusions are statistically sound and interpretable. Together these guidelines describe the reasoning a sponsor or investigator must document to justify why a trial is designed the way it is. For a research administrator, this area covers protocol-level design choices - randomization schemes, control arms, adaptive design elements, endpoint selection, and statistical analysis plans - that determine downstream requirements for data collection, monitoring, and regulatory review.
Guides
Tornado Diagrams for One-Way Sensitivity Analysis in Health-Economic Models
Tornado diagrams rank one-way deterministic sensitivity analysis results by impact, showing which model inputs a health-economic decision model is most sensitive to — and why they cannot substitute for probabilistic sensitivity analysis.
Cheng-Prusoff Equation: Converting IC50 to Ki
The Cheng-Prusoff equation converts an assay-dependent IC50 into an assay-independent Ki. Covers the derivation logic, a worked numeric example, and the conditions (non-competitive inhibition, allosteric mechanisms, unaccounted substrate concentration, tight-binding inhibitors) that make the conversion invalid.
Expected Value of Perfect Information (EVPI) in Health-Economic Decision Analysis
EVPI quantifies what eliminating all parameter uncertainty in a cost-effectiveness model would be worth, in decision-relevant monetary terms. This guide covers how it is computed from a probabilistic sensitivity analysis, how population-level EVPI (population size x time horizon) is used to judge whether further research is worth funding, and how EVPPI refines the question to which specific parameters are worth resolving.
Expansion Cohort Trial Design in Oncology Trials
How expansion cohorts extend a Phase 1 oncology first-in-human trial into seamless, multi-cohort Phase 2 testing under one IND, and how the design differs from basket and umbrella trials.
AI Reporting Guidelines: CONSORT-AI, TRIPOD+AI, DECIDE-AI and the Rest
A decision guide to the AI-specific reporting-guideline family: which of CONSORT-AI, SPIRIT-AI, TRIPOD+AI, DECIDE-AI and CLAIM applies to your study stage, what each covers, and where to find the checklists.
STARD, STARD-AI and TRIPOD+AI: Reporting Diagnostic Accuracy and Prediction Studies
A working comparison of the STARD 2015, STARD-AI, and TRIPOD+AI reporting checklists for diagnostic-accuracy and AI/ML prediction-model studies: what each covers, how many items, when to use one, more than one, or all three, and how to submit them to a journal.
SPIRIT 2025: Writing a Clinical Trial Protocol (The Full Checklist)
SPIRIT is the reporting checklist for clinical trial protocols — the CONSORT counterpart for the protocol stage. This guide covers the SPIRIT 2025 statement’s 34-item checklist by section, what changed from SPIRIT 2013, who requires it, and how it fits with CONSORT and ICH M11.
STROBE Checklist for Observational Studies: The Full 22-Item Guideline
The complete 22-item STROBE checklist for reporting cohort, case-control, and cross-sectional studies, organized by section, with the four items that differ by design.
Statistical Analysis Plan Template for Clinical Trials
A standard, section-by-section statistical analysis plan (SAP) template for clinical trials, showing exactly where ICH E9(R1) estimand attributes, analysis populations, missing-data handling, and interim-analysis rules belong.
Cluster Randomised Trials and the Intracluster Correlation Coefficient (ICC)
What the intracluster correlation coefficient (ICC) measures in a cluster randomised trial, how it produces the design effect that inflates required sample size, why the number of clusters matters more than total participants, how to analyse clustered data correctly, and what the CONSORT cluster extension requires you to report.
Interventional Studies: Definition, Regulatory Classification, and Design Types
What makes a study interventional rather than observational, how the ICMJE, NIH, FDA, and EU Clinical Trials Regulation define the distinction, intervention types, design families, and registration obligations.
3+3 Design vs. Continual Reassessment Method (CRM) in Phase 1 Oncology Dose-Escalation
A statistical comparison of the 3+3 rule-based dose-escalation design and the model-based Continual Reassessment Method (CRM) used to find the maximum tolerated dose in Phase 1 oncology trials.
Digital Biomarkers and Digital Endpoints in Clinical Trials: Validation and Regulatory Acceptance
A practical guide for research administrators and sponsors to the validation chain that turns raw sensor data into a regulatory-grade digital endpoint, and the FDA pathways that can qualify a digital biomarker for reuse across trials.
Case-Control Study: Design, Odds Ratios, and Common Pitfalls
A practical guide to case-control study design: selecting cases and controls, measuring exposure retrospectively, calculating odds ratios, avoiding recall and selection bias, nested and case-cohort variants, STROBE reporting, and the IRB/consent and HIPAA considerations for chart-review and secondary-data case-control studies.
Cohort Study: Design, Types, and How It Works
A guide to cohort study design in clinical and epidemiological research: prospective vs. retrospective cohorts, how cohort studies compare to case-control, cross-sectional, and RCT designs, key measures (incidence, relative risk, hazard ratio), and their core strengths and limitations.
Hybrid Effectiveness-Implementation Trial Designs Explained
Hybrid effectiveness-implementation designs test whether an intervention works and how to implement it in the same study. This guide explains the Curran et al. typology (Type I, II, III), when each fits, and what it means for protocol design, outcomes, and staffing.
Stepped-Wedge and Cluster-Randomized Trial Designs: When Sponsors Use Them and Regulatory/Ethical Considerations
A practical guide for research administrators and IRBs to stepped-wedge and cluster-randomized trial designs: why sponsors choose them, how they differ from individually randomized trials, and the CONSORT, Common Rule, and Ottawa Statement considerations that apply.
Target Trial Emulation: Applying RCT Design Principles to Real-World Data
What target trial emulation is, the seven-component Hernán & Robins protocol framework, the biases it addresses (immortal time bias, confounding by indication), and how it differs from digital twin synthetic control arms.
ICH E11A Pediatric Extrapolation Guideline: What Sponsors Need to Know
ICH E11A gives sponsors a harmonized framework for using adult or older-pediatric data to support conclusions in younger pediatric populations. Covers what it means, how it relates to E11(R1), and FDA/EMA/NMPA adoption status.
ICH E20: The Harmonized Guideline on Adaptive Designs for Clinical Trials
ICH E20 is the draft ICH-harmonized guideline on adaptive designs for confirmatory clinical trials — currently at Step 2b. This guide covers its scope, current status, and how it relates to FDA’s 2019 adaptive-design guidance and other ICH guidelines.
ICH E17: General Principles for Planning and Design of Multi-Regional Clinical Trials
ICH E17 governs how sponsors plan and design a single clinical trial protocol usable for regulatory submission across multiple regions at once. Adoption timeline, core planning principles, and how it relates to ICH E5, E6, E8(R1), and E9.
FDA’s Project Optimus: Oncology Dose Optimization Explained
FDA’s Project Optimus reforms oncology dose selection, moving sponsors away from maximum-tolerated-dose-only strategies toward randomized dose optimization backed by PK/PD, safety, and efficacy evidence.
CONSORT Checklist for RCT Reporting: The Full 30-Item Guideline
The full CONSORT 2025 30-item checklist reproduced as a table, a copyable flow diagram template, an explicit CONSORT 2010-to-2025 comparison, and the journal submission workflow.
Designing a Clinical Trial Protocol: Structure, ICH E6/E8, and Registration
A practical guide to designing a clinical trial protocol: the core sections ICH E6(R2) and E8(R1) expect, how the protocol relates to the SAP, safety monitoring, and ClinicalTrials.gov registration.
Clinical Outcome Assessment Validation: Context of Use, Measurement Properties, and FDA Qualification
How clinical outcome assessments are actually validated: context of use, the four measurement properties FDA evaluates (content validity, construct validity, reliability, ability to detect change), and how validation differs from formal COA qualification.
Digital Twins in Clinical Trials: Synthetic Control Arms and FDA’s Evolving Posture
What an AI-generated digital twin is, how it’s used to shrink a real placebo/control arm via covariate adjustment, the Unlearn.AI/PROCOVA precedent, and FDA’s evolving (not yet settled) regulatory posture.
Surrogate Endpoint Validation in Clinical Trials: The Regulatory Framework
A research-administrator’s guide to surrogate endpoint validation: biological plausibility, the Prentice criteria, meta-analytic validation, and how FDA’s accelerated approval pathway relies on validated vs. reasonably likely surrogates.
Adaptive Design in Clinical Trials: Types, FDA Guidance & Key Considerations
A guide to adaptive clinical trial design: what qualifies as an adaptive design, common adaptation types (sample size re-estimation, dropping arms, response-adaptive randomization, seamless Phase 2/3), the FDA’s December 2019 adaptive design guidance, and the statistical/operational safeguards that keep an adaptive trial defensible.
Designing a Clinical Trial: Endpoints, Sample Size, Randomization, SAP
A practical guide to the core mechanics of clinical trial design: choosing primary and secondary endpoints, calculating sample size and power, selecting a randomization scheme, and the role of the statistical analysis plan (SAP).
Clinical Study Design: The Major Types and How They Relate
An orientation to clinical study design: observational designs (cohort, case-control, cross-sectional), interventional trials and RCTs, and where adaptive design and decentralized trials fit within the design landscape.
COA Rater Training and Certification: ClinRO and PerfO Requirements
What clinical outcome assessment (COA) rater training and certification involves for ClinRO and PerfO instruments, why inter-rater reliability matters, the regulatory basis under ICH GCP and FDA PFDD guidance, and how this differs from ePRO.
What Makes Oncology Clinical Trials Different: RECIST, Endpoints, and the NCTN
Oncology trials share GCP, IRB, and safety-reporting requirements with every other therapeutic area, but differ in three concrete operational ways: RECIST 1.1 tumor response assessment, an oncology-specific endpoint set (ORR, PFS, OS, DFS), and the NCI’s National Clinical Trials Network cooperative group structure.
Sham Procedure Design in Device Clinical Trials
Sham-controlled device trials use a fake surgical or implantation procedure as the control arm. Unlike a drug placebo, the sham carries real physical risk (anesthesia, incision, implantation) with no prospect of therapeutic benefit, which is why FDA design guidance and IRB/REC risk-benefit review treat it as a distinct case from ordinary placebo control.
N-of-1 Trial Design: Single-Patient Crossover Methodology
N-of-1 trial design uses randomized, blinded, multi-period crossover comparisons within a single patient to determine which treatment works best for that individual, rather than for a population average.
Clinical Trial Phases: FDA Definitions, Milestones, and Trial Administration
FDA’s Phase 0 through Phase IV framework explained: what each phase tests, typical participant counts and duration, the IND/NDA/BLA milestones between phases, and how phase classification shapes staffing, monitoring intensity, and safety-reporting burden.
